D.C. Circuit Denies Stay on Claude AI Ban; May 19 Oral Arguments

A federal appeals court in Washington denied Anthropic’s emergency request to pause the Pentagon’s “supply chain risk” designation blocking its Claude AI models for DoD work. On April 8, the D.C. Circuit said the government can keep the Claude AI ban in place while the case proceeds, despite acknowledging Anthropic would likely suffer irreparable harm, including financial and reputational damage. The court’s balance-of-equities analysis favored the Pentagon. It highlighted how the U.S. manages AI security during an active military conflict. The decision also sets expedited oral arguments for May 19, 2026, which could influence future U.S. AI procurement policy. The dispute began after late-February 2026 negotiation breakdowns over guardrails. The Pentagon objected to Anthropic’s proposed terms that, it said, did not sufficiently remove restrictions on (1) fully autonomous weapons systems (including armed drone swarms without human oversight) and (2) mass surveillance of U.S. citizens. Anthropic argued it offered limited, case-by-case exceptions but refused to drop core safety protections. After President Trump directed federal agencies to stop using Anthropic, Defense Secretary Pete Hegseth issued the supply chain risk designation. Pentagon contractors—including Amazon, Microsoft, and Palantir—were required to stop using Claude for defense-related work. Anthropic called the move unlawful and filed parallel suits. A California judge (Rita F. Lin) had previously granted a preliminary injunction that temporarily lifted enforcement, but the D.C. Circuit’s April 8 order reverses that outcome, creating legal uncertainty until the merits are decided. Traders should note the ruling preserves the status quo: the Claude AI ban remains active during litigation. (Keyword focus: Claude AI ban; supply chain risk designation.)